;;; Hey, emacs(1), this is -*- Mode: TDL; Package: LKB; -*- got it? ;;; ;;; this file was adopted for /ntnu/noen from the /uio/noen/ ;;; 11-03 ;;; db ;;; ;;; _fix_me_ ;;; the following should be able to either condition on a specific ARG2 value ;;; of type `x' or use a FILTER component with type `a', so as to only add the ;;; selected preposition EP when there is an actual argument. (1-feb-04; oe) ;;; careful_j_about := monotonic_mtr & [ CONTEXT.RELS < [ PRED "_careful_j_rel", LBL #h1, ARG2 #x1 & [ MARK test ] ] >, FILTER.RELS < [ PRED "_careful_j_rel", LBL #h1, ARG2 a ] >, OUTPUT.RELS < [ PRED _about_p_sel_rel, LBL #h1, ARG0 e_nontense, ARG2 #x1 & [ MARK complete ] ] > ]. ;;; ;;; _fix_me_ ;;; for rules of this type not to inflate ambiguity, we would like the spurious ;;; derivation filter (based on derivation overlap and MRS equivalence) to be ;;; available first. (15-feb-04; oe) ;;; ;;; --- which (fix for PR transfer/17) we think we have now. (2-mar-04; oe) ;;; ;;; _fix_me_ ;;; however, we are left with two problems: (a) this rule will also apply to ;;; EPs that have no ARG0 (e.g. messages; fix requires +subsumes+ operator) and ;;; (b) we seem to output results where the rule could have applied further; i ;;; fail to see the reason for the latter; return to this later. (2-mar-04; oe) ;;; ;;; --- at least the latter should be resolved as of yesterday. (7-mar-04; oe) ;; #| perf_default := monotonic_mtr & [ INPUT.RELS < [ ARG0 [ PERF na ] ] >, OUTPUT.RELS < +copy+ & [ ARG0 [ PERF - ] ] > ]. |# ;;; ;;; while the ERG has no ARG0 on adverbs, see whether we can re-arrange things ;;; to still get topicalization right. ;;; ;;; _fix_me_ ;;; this rule nicely showcases a range of deficiencies in our current set-up: ;;; (a) it would be nice to use a coreference for the PRED of the message (PR ;;; transfer/63) or the +copy operator (which does not allow overwriting the ;;; TPC value, however) and (b) we would like to be able to refer to the set of ;;; adverbs (with respect to the ERG SEM-I) that need their ARG0 ditched. ;;; (7-mar-04; oe) adverb_topicalization := monotonic_mtr & [ INPUT.RELS < [ PRED prpstn_m_rel, LBL #h1, MARG #h2, TPC #i3 ], [ PRED "_alternatively_r_rel", LBL #h4, ARG0 #i3, ARG1 #e5 & [ MARK test ] ] >, OUTPUT.RELS < [ PRED prpstn_m_rel, LBL #h1, MARG #h2, TPC #e5 ], [ PRED "_alternatively_r_rel", LBL #h4, ARG0 [ DITCH + ], ARG1 #e5 & [ MARK complete ] ] > ].